Daniel Ribeiro <drwyrm@xxxxxxxxx> writes:
> The UDC may be connected to an external USB transceiver on SPI or I2c, this
> patch allows setting the pxa2xx_udc parent to assert proper suspend/resume
> ordering.
>
> Signed-off-by: Daniel Ribeiro <drwyrm@xxxxxxxxx>
>
> ---
> arch/arm/mach-pxa/devices.c | 15 ++++++++++-----
> arch/arm/mach-pxa/include/mach/udc.h | 1 +
> arch/arm/mach-pxa/pxa25x.c | 1 -
> arch/arm/mach-pxa/pxa27x.c | 1 -
> 4 files changed, 11 insertions(+), 7 deletions(-)
>
> diff --git a/arch/arm/mach-pxa/devices.c b/arch/arm/mach-pxa/devices.c
> index d245e59..db0228d 100644
> --- a/arch/arm/mach-pxa/devices.c
> +++ b/arch/arm/mach-pxa/devices.c
> @@ -15,6 +15,7 @@
> #include <mach/camera.h>
> #include <mach/audio.h>
> #include <mach/pxa3xx_nand.h>
> +#include <mach/hardware.h>
>
> #include "devices.h"
> #include "generic.h"
> @@ -71,12 +72,18 @@ void __init pxa_set_mci_info(struct pxamci_platform_data *info)
> pxa_register_device(&pxa_device_mci, info);
> }
>
> -
> -static struct pxa2xx_udc_mach_info pxa_udc_info;
> +void __init pxa_set_udc_parent(struct device *parent_dev)
> +{
> + pxa25x_device_udc.dev.parent = parent_dev;
> + pxa27x_device_udc.dev.parent = parent_dev;
> +}
>
> void __init pxa_set_udc_info(struct pxa2xx_udc_mach_info *info)
> {
> - memcpy(&pxa_udc_info, info, sizeof *info);
> + if (cpu_is_pxa27x())
> + pxa_register_device(&pxa27x_device_udc, info);
> + else if (cpu_is_pxa25x())
> + pxa_register_device(&pxa25x_device_udc, info);
> }
>
> static struct resource pxa2xx_udc_resources[] = {
> @@ -100,7 +107,6 @@ struct platform_device pxa25x_device_udc = {
> .resource = pxa2xx_udc_resources,
> .num_resources = ARRAY_SIZE(pxa2xx_udc_resources),
> .dev = {
> - .platform_data = &pxa_udc_info,
> .dma_mask = &udc_dma_mask,
> }
> };
> @@ -111,7 +117,6 @@ struct platform_device pxa27x_device_udc = {
> .resource = pxa2xx_udc_resources,
> .num_resources = ARRAY_SIZE(pxa2xx_udc_resources),
> .dev = {
> - .platform_data = &pxa_udc_info,
> .dma_mask = &udc_dma_mask,
> }
> };
> diff --git a/arch/arm/mach-pxa/include/mach/udc.h b/arch/arm/mach-pxa/include/mach/udc.h
> index 2f82332..6378f30 100644
> --- a/arch/arm/mach-pxa/include/mach/udc.h
> +++ b/arch/arm/mach-pxa/include/mach/udc.h
> @@ -5,4 +5,5 @@
> #include <asm/mach/udc_pxa2xx.h>
>
> extern void pxa_set_udc_info(struct pxa2xx_udc_mach_info *info);
> +extern void pxa_set_udc_parent(struct device *parent_dev);
>
> diff --git a/arch/arm/mach-pxa/pxa25x.c b/arch/arm/mach-pxa/pxa25x.c
> index 77c2693..228ed74 100644
> --- a/arch/arm/mach-pxa/pxa25x.c
> +++ b/arch/arm/mach-pxa/pxa25x.c
> @@ -321,7 +321,6 @@ void __init pxa26x_init_irq(void)
> #endif
>
> static struct platform_device *pxa25x_devices[] __initdata = {
> - &pxa25x_device_udc,
> &pxa_device_ffuart,
> &pxa_device_btuart,
> &pxa_device_stuart,
> diff --git a/arch/arm/mach-pxa/pxa27x.c b/arch/arm/mach-pxa/pxa27x.c
> index a425ec7..7dd8bcf 100644
> --- a/arch/arm/mach-pxa/pxa27x.c
> +++ b/arch/arm/mach-pxa/pxa27x.c
> @@ -346,7 +346,6 @@ void __init pxa27x_set_i2c_power_info(struct i2c_pxa_platform_data *info)
> }
>
> static struct platform_device *devices[] __initdata = {
> - &pxa27x_device_udc,
> &pxa_device_ffuart,
> &pxa_device_btuart,
> &pxa_device_stuart,
> --
> tg: (577c9c4..) pxa/set_udc_parent (depends on: master)
> total: 0 errors, 0 warnings, 61 lines checked
>
> pxa_set_udc_parent.patch has no obvious style problems and is ready for submission.
I cross-posted that to linux-usb mailing list, where IMO this matter should be
also discussed. Just for my understanding, would you explain to me why the order
is important ?
My naive view was that when suspend is triggered, you have either :
(1)
- pxa27x_udc suspend() called => udc is disabled
- transceiver suspend() called => lines are disconnected
(2)
- transceiver suspend() called => line are disconnected
=> pxa27x_udc doesn't receive anymore date
=> packet loss, but that's acceptable since we disable udc anyway
- pxa27x_udc suspend() called => udc is disabled
I don't see why order is important in that case. I understand the conceptual
idea, I just don't understand why.
Cheers.
--
Robert
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
